Conditions for TaIV–TaIV Bonding in Trirutile LixMTa2O6 - Inorganic

Jan 29, 2015 - Investigation of Li insertion into the trirutile structure of MTa2O6 with M = Mg, Cr, Fe, Co, and Ni indicates that Ta–Ta bonding acr...
0 downloads 0 Views 3MB Size
Article pubs.acs.org/IC

Conditions for TaIV−TaIV Bonding in Trirutile LixMTa2O6

Asha Gupta,† Preetam Singh,† Hugo Celio,† C. Buddie Mullins,†,‡ and John. B. Goodenough*,† †

Texas Materials Institute and ‡McKetta Department of Chemical Engineering and Department of Chemistry, The University of Texas at Austin, Austin, Texas 78712, United States S Supporting Information *

ABSTRACT: Stabilization of Ta−Ta bonding in an oxide across a shared octahedral-site edge of a Ta2 dimer is not known. Investigation of Li insertion into the trirutile structure of MTa2O6 with M = Mg, Cr, Fe, Co, and Ni indicates that Ta−Ta bonding across the shared octahedral-site edge of the dimer can be stabilized by a reversible electrochemical reduction of TaV to TaIV for M = Cr, Fe, Co, and Ni but not for M = Mg. Chemical reduction of MTa2O6 by n-butyl lithium only reduced NiTa2O6 to any significant extent. With M = Fe, Co, or Ni, electrochemical formation of the Ta−Ta bonds is accompanied by a partial reduction of the FeII, CoII, or NiII to Fe0, Co0, or Ni0. For M = Cr, two Li per formula unit can be inserted reversibly with no displacement of Cr0. For M = Mg, no MgII are displaced by Li insertion, but a solid−electrolyte interphase (SEI) layer is formed on the oxide with no evidence of Ta−Ta bonding. Stabilization of Ta−Ta bonding across a shared octahedral-site edge in a dimer appears to require significant hybridization of the TaV 5d0 and M 4s0 states.



c-axis bonding. Moreover, the Cr2+ 3d4−Ta5+ 5d0 interaction, which lowers the CrII energy to where it becomes accessible5 in CrTa2O6, does not raise the Ta5+ 5d0 to where it is inaccessible electrochemically in CrTa2O6. Moreover, Ta−Ta bonding occurs in CrTa2O6 without displacement of Cr0. However, only NiTa2O6 could be reduced sufficiently by n-butyl lithium (nBuLi) for X-ray diffraction to show direct evidence of Ta−Ta bonding.

INTRODUCTION We investigate under what conditions reduction of TaV to TaIV in an oxide Ta2 dimer may occur where the Ta 5d electrons are stabilized by Ta−Ta bonding across a shared octahedral site. TaO2 crystallizes in the rutile structure with a shared octahedralsite edge having a Ta−Ta bond distance of 3.018 Å.1 This observation invites investigation of whether, and under what conditions, Ta−Ta bonding across a shared octahedral-site edge of a Ta2 dimer can be stabilized in an oxide. The trirutile MTa2O6 compounds permit an experimental investigation of this question. The role of Ta−O:2pπ−Ta interactions differ in a dimer from those in TaO2. The structure and magnetic interactions of the trirutile oxides have been the subject of investigations for many years.2−4 The MTa2O6 trirutile structure with space group P42/mnm for M = Fe Co, Ni, Mg and P21/n for M = Cr (Figure 1) contains c-axis chains of edge-shared octahedra as in TiO2 rutile, but the cations are ordered into Ta−Ta pairs alternating with single M-atom octahedra in a chain. The counter cations M = Cr, Fe, Co, or Ni provide empty 4s states that are significantly more stable with a larger radial extension than the 3s states of Mg, and the energy of the 4s states can be expected to be close enough to that of the lowest Ta 5d states for strong hybridization. Moreover, the rutile structure allows Li diffusion in the empty octahedral sites along the c-axis to give a room-temperature reversible electrochemical reduction of the cations by Li insertion. A reversible insertion (after the initial formation of a passivating SEI surface layer in the initial Li-insertion reaction) of Li would signal reduction of the cations in situ rather than a displacement reduction reaction of the TaV to Ta0. Our experiments show a reversible reaction of 3 to 4 Li atoms per formula unit (fu) for M = Cr, Fe, Co, or Ni but not for M = Mg, showing that strong hybridization of M 4s and Ta 5d states, but not of Mg 3s and Ta 5d states, stabilizes Ta−Ta © XXXX American Chemical Society



EXPERIMENTAL SECTION

Powder samples of MTa2O6 (M = Mg, Cr, Fe, Co, Ni) were synthesized by conventional solid-state reaction as summarized in Table 1. The precursor oxides were mixed for at least 60 min with an agate mortar and pestle and pressed into pellets if required; the heating rate was kept at 3 °C/min. In the case of FeTa2O6, heating was carried out under an argon 5% H2 atmosphere to keep FeII in its reduced state, and in the case of CrTa2O6, samples were heated in an evacuated sealed quartz tube. Electrochemical Li half-cell tests were carried out in standard CR2032 coin cells. The cathode was a well-mixed composite of MTa2O6, Super P carbon black conductor, and polytetrafluoroethylene (PTFE) binder in a mass ratio of 70:25:5. The composite electrodes were rolled into a thin sheet and punched into circular disks having a typical mass of 4 ± 2 mg. The electrolyte blend, procured from BASF, consisted of a 1 M solution of LiPF6 in 1:1 (v/v) of ethylene carbonate (EC)/diethylene carbonate (DEC). The cells were assembled and sealed in an argon-filled glovebox; they were allowed to age for 12 h to ensure full absorption of the electrolyte into the composite electrode before they were taken out of the glovebox to an Arbin battery-testing system. The half-cells were cycled at room temperature with a 2 min rest period after each charge and discharge. The chemical lithium insertion was carried out by immersing MTa2O6 in n-BuLi reagent for 24 h with continuous stirring Received: December 9, 2014

A

DOI: 10.1021/ic5029417 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

For X-ray photoelectron spectroscopy (XPS), the coin cells were cycled to a desired potential; they were then dismantled and opened inside an argon-filled glovebox where they were washed with anhydrous dimethyl carbonate (DMC) several times to remove EC/DEC and LiPF6. The samples were then taken out of the glovebox and transferred into the ultrahigh vacuum XPS chamber without exposing them to air by employing a sample transfer interface. This interface was built at the Surface Analysis Laboratory of the Texas Materials Institute (TMI); its design contains a set of built-in figures of merit that were used to verify that samples were not exposed to additional traces of oxygen and water during transfer. Details of the design of this transfer interface will be published elsewhere. XPS was performed with a Kratos Axis Ultra, a multitechnique electron spectrometer (Manchester, U.K.). The monochromatic Al X-ray source (148.6 eV) was calibrated with the 3d5/2 line of silver, while the XPS spectra were calibrated with references to graphitic carbon at 284.5 eV in the electrode or adventitious carbon at 285 eV to check or correct the binding-energy of the peaks due to charging effects. Metallic regions of the XPS spectra were recorded with eight scans for signal averaging, a dwell time of 4500 ms, an aperture slot of 300 × 700 μm, a pass energy of 20 eV, and 0.1 eV per step. The pressure in the analysis chamber was typically 2 × 10−9 Torr during data acquisition. Casa XPS analysis software was used for peak deconvolution, and line syntheses were conducted with a Gaussian−Lorentzian 70:30 curve fit and Shirley background subtraction of elemental spectra.



RESULTS AND DISCUSSION The color of the as-synthesized MTa2O6 varied from black for CrTa2O6 to white for MgTa2O6 depending on the M element present in the oxide (see Table 1). Rietveld refinement of the powder XRD spectra of MTa2O6, Figure 2, was carried out by simultaneously varying the structure factor, background parameters, unit cell, profile parameters, and isotropic thermal parameters. The unit-cell parameters, given in Table 1, are in close agreement with those given in the literature. SEM images in Supporting Information, Figure S1 show that the crystallites of MTa2O6 are highly agglomerated and form micron-sized particles. The first and second half-cell discharge (lithiation) curves of CoTa2O6 at 0.2 C rate are shown in Figure 3a. The plateau of 1.1 V in the region of B to C corresponds to the two-phase reaction

Figure 1. Trirutile tetragonal crystal structure of MTa2O6; the center M2+ is shown with its polyhedron denoting the octahedral coordination. M2+ ions occupy the corner (0, 0, 0) and center (1/2, 1/2, 1/2) of the tetragonal primitive unit cell. The two M2+ sites are nonequivalent because the O2− octahedra that surround each M have their in-plane principal axes rotated by 90° with respect to one another. Ta atoms are at positions (0, 0, ±z) and (1/2, 1/2, ±z), where z is close to 1/3. in an Ar-filled glovebox, followed by extensive washing and drying at 60 °C. Powder X-ray diffraction (XRD) spectra were obtained with a Philips X’Pert diffractometer and Cu Kα radiation (λ = 1.5418 Å) to ensure the completeness of the reaction and to characterize the room-temperature phase. The angular resolution in 2θ scans was 0.02° over 10° < 2θ < 100° at a scan rate of 0.25° min−1. The XRD patterns were refined by the Rietveld method with the Fullprof-fp2 program.6 The surface morphology of the sintered powders was investigated in a scanning electron microscope (SEM, Quanta FEG650).

3Li0 + CoIITa 2VO6 = Li3Co0.5Ta 2O6 + 0.5Co0

(1)

The small step in the region from C to D is equivalent to a reduction of the remaining CoII to Co0. Further lithiation beyond

Table 1 target compound (impurities detected)

precursors, synthesis, or sintering temperature and conditions

CrTa2O6 (little Ta2O5 detected)

Cr2O3, Cr, and Ta2O5; pelletized and sealed in an evacuated quartz tube, 1050 °C, 48 h.7

FeTa2O6

Fe2O3 and Ta2O5; ground powder heated in 5% H2/Ar-atmosphere at 1000 °C for 36h.

lattice parameters (Å) a = 4.7334(5) b = 4.7311(2) c = 9.2733(1); β = 90.1065 a = 4.7584(2)

color black

orange− brown

8

CoTa2O6 (trace amount of unidentified impurity) NiTa2O6

MgTa2O6 (trace amount of Mg4Ta2O9)

c = 9.2038(3) Co3O4 and Ta2O5; ground powder heated 800 °C for 12 h followed by 1050 °C for 48 h with a = 4.7385(2) intermediate grinding and pelletization. c = 9.1758(3)9 NiO and Ta2O5; pelletized and heated at 1400 °C, 48 h. a = 4.7216(3) (MgCO3)4 · Mg(OH)2 · 5H2O and Ta2O5; ground powder heated to 1120 °C, 36 h.

c = 9.1280(7)10 a = 4.7185(6)

pink

yellow− green white

c = 9.2095(4)11 B

DOI: 10.1021/ic5029417 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 2. Observed (red dots), calculated (black line), and difference XRD pattern (blue line) collected from (a) CrTa2O6, (b) FeTa2O6, (c) CoTa2O6, (d) NiTa2O6, and (e) MgTa2O6 at room temperature. The lower green tick marks indicate allowed Bragg reflections from the trirutile tetragonal phase.

insertion/deinsertion reaction. The cyclic voltammogram for this reversible oxidation and reduction process is given in Supporting Information, Figure S2. NiTa2O6 behaves similarly to CoTa2O6 as seen in Figure 3b. The B−C plateau appears at 1.25 V and corresponds to the twophase reaction similar to CoTa2O6.

D is associated with formation of a solid−electrolyte interphase (SEI) layer; beyond the region from E to F the sample surface becomes increasingly amorphous, as is seen in the XRD of Figure 4a. After the first discharge to F, the B−C plateau disappears, and the reversible capacity drops. Repeated cycling of over three Li+ ions per formula unit (fu), Figure 4b, indicates that the Co0 and TaIV are reoxidized to CoII and TaV in the CoTa2O6 structure within an amorphous mass that transports Li+ in the Li+-ion

3Li0 + Ni IITa 2VO6 = Li3Ni 0.5Ta 2O6 + 0.5Ni0 C

(2)

DOI: 10.1021/ic5029417 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 3. (a−e) Lithiation/delithiation behavior of MTa2O6 at a 0.2 C cycling rate. Points A−G denote the state corresponding to the XRD data in Figure 4a and Supporting Information, Figure S3.

The charge/discharge curves of CrTa2O6 and FeTa2O6 show similar behavior but are different from those of CoTa2O6 and NiTa2O6. Figure 3c shows the first and second half-cell discharge (lithiation) of FeTa2O6. The large B−C plateau of Figure 3a for CoTa2O6 was dramatically shortened, while the pseudo C−D plateau has been increased to over four Li atoms per fu, indicating that the reduction reaction from FeII to Fe0 does not have a step at 0.5 Fe0 as occurs in the lithiation of CoTa2O6 because it occurs simultaneously with SEI-layer formation. After the step, the voltage drops from 1.1 to 0.9 V where formation of the SEI layer forms rapidly.

Further lithiation below 1 V leads to formation of the SEI layer; and after the first discharge to point E followed by charging to point F, the reversibility of the B−C plateau disappears. The XRD profiles of the composite electrodes, Supporting Information, Figure S3a, show that at point D, NiTa2O6 starts to lose its crystallinity, and at point F, NiTa2O6 becomes amorphous. The plot of cycle number versus specific capacity, see Supporting Information, Figure S4a, shows that after the first initial irreversible loss due to formation of the SEI layer, the reduced Ni0 and TaIV are reoxidized back to NiII and TaV and participate in the electrochemical Li+ ion insertion/deinsertion reaction. D

DOI: 10.1021/ic5029417 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

stabilized in the absence of cation vacancies in CrTa2O6. The energy of the bottom of the MII 4s0 band decreases with increasing atomic number, which increases the discharge voltage with MTa2O6 as the cathode of a cell with a Li anode. The energy of the TaIV−TaIV bond across a shared octahedral-site edge remains below the bottom of the CrII 4s0 band; it is just overlapped by the FeII 4s0 band, and a stronger overlap of the 4s0 band occurs for M = Co and Ni. In all four compounds, there is a strong hybridization of the TaV 5d0 and MII 4s0 states to enhance the overlap of the TaV 5d0 orbitals across the shared octahedralsite edge. The half-cell first and second discharge curves of Figure 3e for MgTa2O6 show no evidence of a displacement of MgII or of a plateau or pseudoplateau indicative of stabilization of a phase with Ta−Ta bonding. Only 4s states of the transition-metal ions are capable of undergoing a hybridization with TaV 5d states in a manner that assists reduction of TaV to TaIV; the alkaline metal MgII is not capable of aiding the reduction of TaV to TaIV. To determine the cation oxidation states at different states of Li insertion, we recorded the XPS spectra of LixCoTa2O6 at the discharge states A, D, and F of Figure 3a and after the first recharge to G. The core-level binding energy spectra of the Co 2p and Ta 4f atoms are shown in Figure 5. The peak full width at half-maximum was kept constant for the pair of Co 2p peaks at each state of charge, and the ratio of the areas of the 2p3/2 and 2p1/2 peaks was fixed at 2:1. As indicated in Figure 5a for the charged states A and G, the Co 2p spectrum can be fitted with 2p3/2 and 2p1/2 binding energies of 780.4 and 796.2 eV; the two smaller peaks at 786.4 and 802.5 eV are shakeup satellites. The energy difference of ∼16 eV between the 2p3/2 and 2p1/2 peaks corresponds to that of the literature CoII oxidation state12 and also to the pink color of the CoTa2O6 powder. After discharge to point D of Figure 3a, all the Co ions are expected to be reduced to Co0 if there is no capture of Li in an SEI layer on the oxide surface. The Co 2p spectrum for point D shows a coexistence of CoII and Co0 and therefore the existence of an SEI layer. This possibility is confirmed in the test shown in Supporting Information, Figure S5. However, amorphization of the sample appears to occur with insertion of more than three Li+/fu, indicating instability of the structure with the loss of more than 0.5 M/fu. At point G of Figure 3a on delithation, all the reduced Co0 are oxidized back to CoII. The core-level Ta 4f spectra, Figure 5b, of CoTa2O6 at points A and G of Figure 3a consist of two peaks split by 2 eV with binding energies of 26.2 and 28.1 eV that correspond to the Ta 4f7/2 and Ta 4f5/2 of a TaV oxidation state.13 The Ta 4f spectra at points D and F of Figure 3a are split by 1.8 eV, and a small shift toward higher binding energy was observed rather than to lower binding energy as expected for a cation reduction. This kind of anomalous negative binding energy shift is not an uncommon phenomenon, although a physical explanation for this kind of behavior is not clear.14−16 The Ni 2p3/2 and 2p1/2 binding energy peak, Figure 6a, for the starting material NiTa2O6 is observed at 854.1 and 871.4 eV, respectively, and a strong satellite peak is apparent at 860.2 eV (at a distance of ∼6.1 eV from the main 2p3/2 peak), consistent with Ni in a +2 oxidation state17 as expected for NiTa2O6. XPS spectra for Ni 2p recorded for samples at points B−F of the discharge curve of Figure 3b becomes complicated for analysis owing to interference from fluorine KLL-auger peaks with the strongest peak at 858 eV. Fluorine is present both in the polytetrafluoroethylene binder used to make the composite cathode and in the LiPF6 electrolyte; the fluorine from this source deposits on the

Figure 4. (a) XRD patterns of CoTa2O6 in the lithiation process; points A−E correspond to the states indicated in Figure 3a. (b) Discharge voltage profiles of CoTa2O6 at different rates. (inset) The cycle number vs specific capacity.

4Li0 + Fe IITa 2VO6 = Li3Ta 2O6 + 0.5Fe 0 + Li+(in SEI) (3)

The XRD data of Supporting Information, Figure S3b show that reaction with the electrolyte in the formation of an amorphous SEI layer has taken place in the interval C−D in the case of FeTa2O6. Nevertheless, the Fe0 and the TaIV are reoxidized in subsequent charge/discharge cycles; but as in CoTa2O6, the reversible capacity decreases with cycle number and rate, Supporting Information, Figure S4b. Figure 3d shows the first and second half-cell discharge (lithiation) of CrTa2O6; the region of A−C corresponds to the insertion of 2 Li+/fu. 2Li0 + Cr IITa 2VO6 = Li 2Cr IITa 2IVO6

(4)

The XRD data of Supporting Information, Figure S3c show that CrTa2O6 remains crystalline even after one complete discharge and charge cycle. At points D and E, the major diffraction peaks are still detectable. This observation along with the reversibility of the consecutive cycle (see Supporting Information, Figure S4c) indicates that unlike Fe, Co, and Ni, CrII is not pushed out of the lattice as Cr0, and the lattice structure remains intact after the insertion of two Li atoms. Reduction of the TaV to TaIV without an accompanying displacement of Cr0 shows that formation of a TaIV−TaIV bond across a shared octahedral-site edge may be E

DOI: 10.1021/ic5029417 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 5. XPS spectra of (a) Co 2p; the black dots represent the observed data, the solid red and blue lines correspond to Co 2p3/2 and Co 2p1/2 states, and the dotted red and blue lines represent the satellite peaks; the bold green envelope is the fitted curve.(b) Ta 4f. A, D, F and G correspond to the charge/discharge states indicated in Figure 3a.

Figure 6. XPS spectra of (a) Ni 2p from NiTa2O6 and (b) Fe 2p from FeTa2O6. A−F correspond to the charge/discharge states indicated in Figure 3b,c, respectively. The vertical dashed line corresponds to the binding energy position of Fe2+, while the arrow indicates the position at which Fe0 is expected.

for NiTa2O6 at the different states of charge display an evolution similar to those for CoTa2O6. Figure 6b shows the XPS spectra of the Fe 2p at points A, C, D, and E of the discharge curve of Figure 3c and at point F at the end of the first charge. At point A of the as-synthesized FeTa2O6, the 2p3/2 and 2p1/2 peaks appear at 709.7 and 723 eV, respectively, with a shakeup satellite peak ∼5 eV from the main 2p3/2 peak.18 The satellite of the 2p1/2 is completely obscured by interference from a broad F 1s plasmon loss from the Teflon-like PTFE binder, see Supporting Information, Figure S6. On cycling, changes in the sample binder shift the energy of the F 1s plasmon loss to lower energy. Nevertheless, the Fe 2p2/3 peak is not completely obscured, and the intensity of the FeII 2p3/2 peak is

NiTa2O6 crystallites during the formation of the SEI layer. Because of the dynamic process of the electrochemical cycling, the amount of fluorine and its interference in the Ni 2p spectra is not stable and varies for all the spectra recorded in the B−F region of Figure 3b. Nevertheless, we could still gauge the positions of the Ni 2p peaks in their two oxidation states, namely, +2 and 0. As the process of lithiation progresses from point B to E, the intensity of Ni2+ 2p3/2 decreases, while a peak at 852 eV due to Ni0 2p3/2 keeps increasing. At point F of Figure 3b, the corresponding XPS spectra show that the Ni2+ 2p3/2 peak is restored and that the Ni0 2p3/2 disappears. This observation indicates that the NiII and Ni0 participate reversibly in the charging/discharging process of the half-cell. The Ta 4f spectra F

DOI: 10.1021/ic5029417 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Table 2. Comparison of Structural Parameters and Bond Lengths of MTa2O6 and Its Chemically Lithiated Analogue LixMTa2O6 NiTa2O6 LixNiTa2O6 CoTa2O6 LixCoTa2O6

lattice parameters (Å)

unit cell volume (Å3)

Ta−Ta bond along c-axis (Å)

M−Ta bond along c-axis (Å)

Ta−Ta bond along [116] direction (Å)

a = 4.7216(3) c = 9.1280(7) a = 4.7180(4) c = 9.1203(3) a = 4.7385(2) c = 9.1758(3) a = 4.7369(7) c = 9.1727(5)

203.50

3.0962(7)

3.0159(7)

3.6471(4)

203.01

3.0402(3)

3.0398(3)

3.6658(3)

206.03

3.1124(3)

3.0317(3)

3.6611(8)

205.82

3.1114(5)

3.0321(5)

3.6831(2)

Figure 7. Observed (red dots), calculated (black line), and difference XRD pattern (blue line) collected from (a) LixCoTa2O6 and (b) LixNiTa2O6 at room temperature. The lower green tick marks indicate allowed Bragg reflections from the trirutile tetragonal phase.

contribution of a lowering of the Ta 5d orbital energy is stronger in TaO2, where the interactions are active in 3D, than in MTa2O6 where they are only active for a single Ta−Ta interaction. Moreover, electrostatic repulsion between the two TaIV of a dimer is not resisted by a positive charge at the M-atom positions, so the TaIV−TaIV separation is reduced in NiTa2O6 but can be expected to be somewhat larger than that in TaO2.

seen to decrease with increasing lithiation from C to D to E, while an Fe0 2p3/2 peak at ∼706 eV increases. At point F of Figure 3c, the intensity of the FeII 2p3/2 peak is restored, and that of Fe0 disappears. The Ta 4f spectra (see Supporting Information, Figure S7) for FeTa2O6 at the different states of charge display an evolution similar to those for CoTa2O6 showing that TaV is reduced to TaIV during Li intercalation and is oxidized back to TaV when Li de-intercalation occurs. To provide direct evidence for Ta−Ta bonding across the shared octahedral-site edge in reduced MTa2O6, we attempted to reduce the compounds chemically with n-BuLi to obtain samples for the XRD analysis without the carbon used to form the cathodes for electrochemical reduction. Results of the XRD analysis are shown in Table 2 and the XRD profiles in Figure 7. After stirring a sample for 24 h in a solution with n-BuLi, the FeTa2O6 showed no color change, and no change of the structure was observed after refinement; the reducing power of n-BuLi is not strong enough to reduce FeTa2O6. Although the color of the CoTa2O6 sample changed from pink to black, the XRD data showed only a small structural change (see Table 2). However, partial chemical lithiation into the bulk of NiTa2O6 showed a decrease in the volume (203.5 to 203.01 Å3) and a reduction of the c-axis Ta−Ta separation (3.096 to 3.042 Å) as against 3.018 Å in TaO2, with a simultaneous increase of the c-axis Ni−Ta separation and Ta−Ta separation between neighboring axes. These changes point to a c-axis Ta−Ta bonding that increases with the reduction of TaV. The Ta−Ta bonding within a dimer is the result not only of direct overlap of 5d orbitals of the two Ta atoms but also of overlap by covalent mixing through the O 2pπ orbital. This latter



CONCLUSIONS Reduction of Ta2O5 by Li insertion in a displacive reaction from TaV to Ta0 has been reported by Dang et al.19 This reduction is to be distinguished from an in situ reduction of TaV to TaIV with the formation of a Ta−Ta bond across a shared octahedral-site edge in an oxide. In this paper, we have explored the conditions under which an in situ reduction of TaV to TaIV can be stabilized by formation of a Ta−Ta bond across a shared octahedral-site edge in the trirutile oxides MTa2O6. We have found that this stabilization of TaIV does not occur in MgTa2O6, but it is found in MTa2O6 with M = Fe, Co, or Ni with a reduction of the MII to M0 occurring simultaneously with the in situ reduction of TaV to TaIV. In situ reduction of TaV to TaIV occurs in CrTa2O6 without a reduction of CrII to Cr0. Hybridization of the Ta 5d and Mg 3s states is smaller than that of the Ta 5d and the 4s states of Cr, Fe, Co, and Ni. Therefore, we conclude that in situ reduction of TaV to TaIV can occur in an oxide where Ta2 octahedral-site dimers share an edge provided there is a strong hybridization of the Ta 5d states with empty 4s states of a countercation; CrTa2O6 forms Ta−Ta bonds without displacement of Cr0 from the lattice, which shows that removal of the M atoms is not responsible for the stabilization of G

DOI: 10.1021/ic5029417 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry TaIV with the formation of Ta−Ta bonding. Direct evidence for the formation of in situ TaIV−TaIV bonding across a shared octahedral-site edge by XRD could only be shown for NiTa2O6 because of the lower reducing power of n-BuLi compared to that of the electrochemical reduction.



(19) Dang, H. X.; Lin, Y.-M.; Klavetter, K. C.; Cell, T. H.; Heller, A.; Mullins, C. B. ChemElectroChem 2014, 1, 158−164.

ASSOCIATED CONTENT

S Supporting Information *

Scanning electron micrographs of MTa2O6, cyclic voltammogram of CoTa2O6, evolution of XRD during lithiation and delithiation processes and charge−discharge voltage profiles of NiTa2O6, FeTa2O6, and CrTa2O6, XPS spectra of Co 2p, interference of F 1s plasmon loss peaks with those of Fe 2p and Ta 4f for FeTa2O6. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We acknowledge The Robert A. Welch Foundation of Houston, Texas [Grant Nos. F-1066 for J.B.G. and F-1436 for C.B.M.] and the Department of Energy Office of Basic Energy Science Grant No. DE SC00005397.



REFERENCES

(1) Syono, Y.; Kikuchi, M.; Goto, T.; Fukuoka, K. J. Solid State Chem. 1983, 50, 133−137. (2) Takano, M.; Takada, T. Mater. Res. Bull. 1970, 5, 449−454. (3) Kremer, R. K.; Greedan, J. E. J. Solid State Chem. 1988, 73, 579− 582. (4) Kinast, E. J.; dos Santos, C. A.; Schmitt, D.; Isnard, O.; Gusmão, M. A.; da Cunh, J. B. M. J. Alloys Compd. 2010, 491, 41−44. (5) Saes, M.; Raju, N. P.; Greedan, J. E. J. Solid State Chem. 1998, 140, 7−13. (6) Rodriguez-Carvajal, J. Multi-Pattern Rietveld Refinement Program FullProf 2k, version 3.30; Laboratoire Léon Brillouin (CEA-CNRS) CEA/Saclay 9119: Gif-sur-Yvette, France, June 2005. (7) Guillen-Viallet, V.; Marucco, J. F.; M. Ghysel, M. J. Alloys Compd. 2001, 317−318, 127−131. (8) Peters, E.; Mueller Buschbaum, H. Z. Naturforsch., B: J. Chem. Sci. 1995, 50, 712−716. (9) JCPDS Card No. 32−314. (10) Ehrenberg, H.; Wltschek, G.; Rodriguez-Carvajal, J.; Vogt, T. J. Magn. Magn. Mater. 1998, 184, 111−115. (11) Ferrari, C. R.; Hernandes, A. C. J. Eur. Ceram. Soc. 2002, 22, 2101−2105. (12) Briggs, D. and Grant, J. T. Surface Analysis by Auger and X-ray Photoelectron Spectroscopy; IM Publications: Chichester, U.K., 2003. (13) Jagadeesh Chandra, S. V.; Uthanna, S.; Mohan Rao, G. Appl. Surf. Sci. 2008, 254, 1953−1960. (14) Kukuruznyak, D. A.; Moyer, J. G.; Prowse, M. S.; Nguyen, N.; Rehr, J. J.; Ohuchi, F. S. J. Electron Spectrosc. Relat. Phenom. 2006, 150, 282−287. (15) Ziegler, Ch.; Frank, G.; Göpel, W. Z. Phys. B: Condens. Matter 1990, 81, 349−353. (16) Gaarenstroom, S. W.; Winograd, N. J. Chem. Phys. 1977, 67, 3500−3506. (17) Gupta, A.; Waghmare, U. V.; Hegde, M. S. Chem. Mater. 2010, 22, 5184−5198. (18) Gupta, A.; Kumar, A.; Waghmare, U. V.; Hegde, M. S. Chem. Mater. 2009, 21, 4880−4891. H

DOI: 10.1021/ic5029417 Inorg. Chem. XXXX, XXX, XXX−XXX